Background: I'm working on a game that has multiple paths for players to choose from. The choices the players make can lead them to very different endings. There are many bad endings but only one true canon ending. Only one good ending, basically.

Could this be a bad idea for the story of a game? I feel like players might think they just wasted their time whenever they go into a route that leads to a game over. I don't want to make it feel like getting the sole good ending is the main goal and the only thing the player should focus on. I still want to try and make each bad ending interesting, but I'm not sure if that alone would make those bad endings worth playing through. Some examples:

  • some bad endings have foreshadowing to the good ending and help players, so they can figure out the right choice to reach said good ending. Example: you find out in one of the bad endings which character was holding the important item secret from everyone else, allowing you to confront him in another run.
  • some bad endings show how a character reacts when a certain event happens, or the effects of keeping a certain character alive. Example: if you make sure a certain character lives until a certain point, that character later manages to prevent the genocide that ended your run in one of the bad endings.

Side Note: because of the nature of the story, it can only have one good ending, not multiple. Don't ask me why, it would be VERY complicated to explain why that's not an option for me...

  • 19
    This quesiton is in the close vote review queue as "off-topic". I am voting to Leave Open as this is about writing practices and their effect on the consument for a story with multiple endings.
    – Secespitus
    Commented Feb 9, 2018 at 17:40
  • 6
    I am curious about your use of the word canon here. Assuming the story is completely original, why are the bad endings not canon?
    – Octopus
    Commented Feb 9, 2018 at 20:15
  • 4
    You can save scum right? Just like with Choose Your Own Adventure books?
    – Mazura
    Commented Feb 9, 2018 at 20:17
  • 1
    I think the idea is good. Gives the player a reason to play through multiple times. Commented Feb 10, 2018 at 14:10
  • 3
    In Undertale, there's just one way to get to the "pacifist" ending and it involves doing many things right. Commented Feb 10, 2018 at 18:35

13 Answers 13


It depends on your genre - in horror games this can be a very good decision

If you are going for a darker themed game, and your description suggests that you are doing this, then having multiple bad endings is fine. I've played quite a few horror games that were created with the RPG Maker (I am not affiliated with the company) for example where there are lots of bad endings - and I enjoyed all of them. Because you spend most of your time in the game itself: The journey is the destination.

Having only bad endings would probably be a bad idea. This would make the player feel like all of their effort was indeed wasted. But if your game makes it obvious that there are multiple endings and each ending gives information about a possible other ending, one of them being good, then everything is fine. If only you had taken the knife with you that the Antagonist just used to stab you, things might have gone a different route...

The players who enjoy playing a game multiple times will try to follow the advice they got. Others might simply look up a walkthrough that shows them the other endings. I've been on both sides, depending on how much time I would need to spend with the same stuff I've already done in previous runs. The more decisions one can do the better. Spending 10 hours only to click Left instead of Right before the Boss' lair is boring. But deciding whether you take the lamp or the book with you after a few minutes is a completely different matter. As long as these decisions matter that is. Having regular checkpoints, especially before important decisions can help, too. If that is your style, of course. Some games prefer to go a rogue-like style and let the players decisions be final when they are deciding which part of the story they want to play.

Having only one good, hard to achieve, ending is especially good for perfectionist gamers that want to get everything. They will try to get every ending and see which hints they contain that will help them get to 100 %.

Furthermore having only one good ending makes this one feel special to the players who manage to achieve it. It's clearly better than any other ending they managed to get and therefore more valuable in their eyes.

Having bad things happen to your character is not necessarily a bad thing, as long as you give it a positive spin. See my answer to the question Still struggling with character desire, positive vs. negative, hooking readers for a more detailed discussion about why negative feelings resonate more strongly in us humans and how this can be used to give a character a motivation that the reader/player can relate to as they are trying to bring happiness to the character that is suffering. Multiple bad endings mean that there are more reasons to help the character finally achieve his well-deserved piece of peace and happiness.

What you also sometimes see is that a game has maybe two or three bad endings (plus a Game Over that can for example occur if the player dies) and one good ending. And then one special, true ending that you can only get if you for example collect all of the very rare items. This is another possibility. Make it a little bit easier to have a nice ending to please your more casual gamers and make the true ending extra hard to get and thereby extra rewarding. The true ending doesn't have to be completely happy by the way, though this will surely leave the player with the feeling of having completed a horror game.

Having only equally bad endings is possible, but in general you want your player to think: "That went better than the last time. Still bad, but I managed to get further. If I keep this up I will make it to the 'good' ending!" That means you should think about making endings in multiple shades of bad - from "Game Over" over "I am dead at the end" over "Well, I managed to run away, but I didn't realize there was a trap outside that room" over "I managed to nearly get away - if only I had a weapon to fight that final monster" to "I managed to get away and lead a happy life". With the true ending being "The monster was my sister/ my father/ myself all along..."

But your game has to fit this style. You don't want a totally happy and relaxed game where you suddenly realize at the end that your decisions have led to a bad ending. Imagine for example a Pokemon game (again, I am not affiliated with the company) where you basically have nothing to lose from spending some time trying to catch or train your Pokemon - and in the end you realize that there was a time limit of 100 in-game days to get to the end and now you can't beat the Top Four. That would suck.

  • 3
    A good example of "each ending gives information about a possible other ending" is the visual novel Fate/stay Night. It has literally dozens of bad endings, and after each one you get a short scene that gives you hints on where you went wrong and how you can avoid getting the same ending next time round. (Fair warning for OP, in case he wants to try it: it contains NSFW material.)
    – F1Krazy
    Commented Feb 9, 2018 at 17:21
  • 4
    I'd also note, that bad endings can often be more impactful than happy, good endings. 'Yay.... hero deus ex machina'd the bad guy... woo....' then the story is done and player gets over it. Where as a 'bad ending' that really kind of ***s with the player's mind and leaves them saying 'wtf?' for the next week or more has left a much more larger impression on the player. Commented Feb 9, 2018 at 22:10
  • That is a nice write up that really helps understand how beneficial having terrible things, or 'bad endings' can be for your story :). That link may possibly be insightful for the OP as well. Commented Feb 9, 2018 at 22:23
  • @TylerDahle Thanks, I added a paragraph with the link and a little discussion geared towards this question.
    – Secespitus
    Commented Feb 9, 2018 at 22:28
  • 1
    @Mołot Yes, that is correct. Having only equally bad endings is possible, but in general you want your player to think: "That went better than the last time. Still bad, but I managed to get further. If I keep this up I will make it to the 'good' ending!"
    – Secespitus
    Commented Feb 12, 2018 at 16:31

This is actually the classic video game narrative. Consider a game like Super Mario Brothers. There is essentially one good ending --rescuing the princess --and everything else is a bad ending!

To make this work, I would suggest you view your game as one long experience that leads ultimately to the good ending. In order to do this, however, you need to make sure the game is enjoyable to play through multiple times. There are several ways to do that:

If the game is short enough, people might be willing to play through it several times. Or, like in the old-school action-based video games, if the point is the experience, not the story, people might be perfectly happy to play over and over again. For a longer, more narrative game, you could have secret routes that allow experienced players to skip early parts of the game --you can design the game to enable or reveal these routes when you play through certain bad endings. Another technique is to allow save points prior to the player making irrevocable choices. Or, you could make the decision point early, so that the player essentially plays a different game each time through.

  • 4
    +1 for mentioning the replayability aspect. If the game is boring to replay, then not many people will want to do it again to get all the endings. But if the changes that lead to the different endings change the plot enough along the way, or the game itself is the point, then it's good.
    – Bobson
    Commented Feb 12, 2018 at 18:40


If you offer choices to the player, but there is only one that yields positive results, then yes, that is bad design.

I do not write video games, but a lot of Table-top RPGs. I always think of success in multiple levels or factors. This allows for the PCs to go "well shucks, we almost got 'im, but at least we [...]"

This also increases the threads for a sequel...

5-point success scale

Instead of one "happy path" and a bunch of bad ones, try and think of successes based on a scale that goes:

Best/ Good/ Okay/ Bad/ Worse

  • Best: Kill the villain, marry the princess, get the loot
  • Good: Kill villain, get some loot
  • Okay: Villain defeated
  • Bad: Villain escapes
  • Worse: PC defeated, princess killed, loot spent by villagers on Christmas ornaments.

Multi-facet victory

Similar to the previous, but there are multiple levels and different elements that can be used to determine success. From the previous example:

  • Villain: Killed/ Escapes/ Wins
  • Princess: Married/ not married/ killed
  • Loot: A lot/ some/ none

This creates a success matrix where you can fail one thread, but succeed at the other two.


It's actually a Good Idea, especially in games

This kind of structure (one good ending, many bad endings) is very common in the genre of visual novels. These kinds of games are essentially choose your own adventure stories, sometimes including interactive gameplay elements, that often include many ways for your point of view character to fail or die before reaching the conclusion of the story.

Bad Endings can be a potent story-writing tool in this kind of medium. They offer a way to ratchet up the tension of scenes by organically letting the player know more than the characters do (as the player has experienced revelations via the bad endings that they characters have not). They also add weight to the player's decisions as they proceed, by making it clear that their own choices led to the ruin of the point of view character and the other characters in the story.

There are some games where the bad endings are not only interesting, but are actually critical to the plot as a whole. I would recommend playing Nine Hours, Nine Persons, Nine Doors or Virtue's Last Reward to understand how this kind of structure can work. In these games, the information the player learns about in the bad endings re-contextualizes the story and provides critical hints to proceeding to the good ending.

  • 2
    All Fallout Endings Since 1997 - 2015
    – Mazura
    Commented Feb 9, 2018 at 21:34
  • "Fallout 1 to NV - Interesting and varied, giving you the opportunity to decide what kind of character you are and an impact on the final result when given a complex choice. The chance for the player to be the villain or hero or an ending which makes you feel sick inside or joyful depending on your choices, showing what kind of person you are. Giving the player the opportunity to replay the game but making a different choice on faction which gives the ending a whole new meaning. Fallout 4 - Who wears Power Armor and who doesn't." – YouTube comment.
    – Mazura
    Commented Feb 10, 2018 at 2:13
  • There are some people "where the bad endings are not only interesting", they're preferred. The OP would be making a mistake IMO by declaring that those endings lack canonicity inside of the game you just played. But in part two, sure: mutants capturing Shady Sands is not canon as of Fallout 2. But it was if that's what you did.
    – Mazura
    Commented Feb 10, 2018 at 2:13

Premature Ends are Necessary

Let’s say, conservatively, all your choices have only two options and no more. Suppose every path through the game lets you make ten choices, and every possible branch you can take is unique.

You would need to write 2,047 unique branches and 1,024 unique endings. If each screen has about as many words on it as the pages in Wikipedia’s list of longest novels, that would put the number of words you’d need to write in between War and Peace and Romance of the Three Kindgoms.

So, out of necessity, you’ve got to limit the exponential growth of your story branches. Either multiple paths all lead you to the same branch, or most of the choices you could make end the story prematurely. Typically, if you’re ending the story early in order to reduce the number of endings, you’re not going to write four or ten or twenty premature endings off that branch and have the player explore them all to discover they’ve put the game in an unwinnable state; you’re going to say Game Over quickly. And typically, a premature end is a bad end.

Thus, the structure of those old Choose Your Own Adventure books, where pages cost real money, so most of the choices you made led to instant gruesome death.

Early Ends Don’t Have to be Bad Ends

This is really up to your creativity as a writer, but an ending just means you don’t have any more choices to make in the game. Realistically, the protagonist’s goal could be to get out of this mess as quickly and easily as possible, or at least be tempted to take the easy way out. This could be at cross purposes to the player’s curiosity about what will happen if things don’t get better—the route that keeps her in a situation she doesn’t want to be in (and you playing the game) could be the bad end! A clever videogame variant I heard about but didn’t play: you start out infiltrating a terror group, they ask you to do various things to prove your loyalty, and the way you keep playing the missions to the end of the (linear) story is to keep playing along with them long past the point where it’s an act, becoming the Big Bad of the game. Actually stopping them before they kill any more people completes your mission, so it’s a Game Over. Inverting the moral valence of that: her story arc might be that she starts out just wishing for an easy way out, and you have choices that will let her do that and be content and happy, but if you keep playing long enough for her to see what’s going on, she’ll start getting more and more uncomfortable and want to change it.

You Can Merge Branches, Not Just Prune Them

Another way to reduce the branching is that a game can have state. Several different choices about what you do over the weekend can all lead you to go back to your normal routine on Monday, but the game can remember that you’ve had different experiences, and that can lead to different outcomes later.

This is the structure of a lot of life-sim games, where most paths through the game take you to the same place at the same time, and then it determines which ending you got based on all the things you did.

Do too much of this, though, and the player might feel you’ve given them false choices that don’t really matter.

A Bad End is in the Eye of the Beholder

Pretty self-explanatory, What’s good for the protagonist might not be a good outcome for the world, and vice versa. She might change her mind about what she wants. We might think the “best” ending is one that gives her the most compelling moral arc, or where she finds out the unpleasant truth, or where she meets her true love, or where she achieves the most glory, but those might not be the only endings that make her happy. And several different happy endings could be long and developed branches that could go wrong in multiple ways. That gives your game replayability even after the player has “won,” beyond seeing all the deaths.


Yes, I would suggest that you find a different story structure. That one is overused.

  • 2
    Oftentimes CYOA and their ilk would also fudge timelines and cause decisions to change the obvious or immediate events but lead to similar i.e. same outcomes. Commented Feb 10, 2018 at 19:56
  • ...This is a good analysis, but I don't see how it leads to your "no". You have my upvote for delving into the history of the medium and the writing tools people use - but I wish to note that it does not constitute support for telling the asker they need to write a different story.
    – Stackstuck
    Commented Sep 6, 2023 at 21:44

This is a matter of opinion. It depends on the audience you will be delivering the game to for play. There is no rule that every story or every game must have a happy ending. In fact, some highly regarded movies have unhappy endings or endings left up to the imagination.

If you want to be successful with your audience, attempt to measure their preference.

If you want to be loyal to your own creative preference, do it the way you desire.

Since you have control over the development of the game you could make their decisions match the outcome. That would most likely deliver the type of outcome that matches the player.


Due to the nature of your game, I'd say it is a good idea, and many games already do do this. The first one that comes to mind is 999 (9 Persons, 9 Hours, 9 Doors) in which you have to do exactly as you are stating- gain knowledge from past play throughs to get the "true" ending.

Giving one true ending is also especially common in visual novel sort of games. Usually with these sorts of games, especially considering that you're encouraging players to get all the endings, the game is only really completed once completing all the "bad" endings to successfully run the "true" ending. The only catch really is to not make the bad ending too unrewarding, but it sounds as though you are aware of that.

Overall I don't think it's bad game design. A player who is really experienced in these sorts of games could always look up the decision tree and just jump to the true ending. I'd suggest maybe a warning at the beginning of the game that your actions will affect the game outcome, as well as a "skip" feature for text/content the player has already seen. Best of luck!


Very few players will finish multiple walkthroughs, unless the game is specifically designed for being played multiple times. Assuming a "standard" RPG, among the players who actually play to the end, the vast majority will only see one ending. Thus you need to make all endings enjoyable.

So you have 2 choices:

  1. Make all endings enjoyable, and make sure they all give closure. The Witcher 3 does that by showing the player over and over again that if the outcome of a choice was bad, the other choice will lead to a different bad outcome. Subsequently all of the possible game endings are good but quite far from perfect - one is a bit better than the others, tough.

  2. Make finding the perfect ending part of the game, don't make a "standard" RPG. An example is Long Live the Queen, a game which is all about making 40 choices, unless an unwise choice leads to your demise before that - the win condition of the game is to find one of the "good" endings.


I think it is a bad idea and might be frustrating to play if it seems like you always lose and cannot ever find the combination that wins.

However, you might be able to make some of those bad choices that lead to certain failure just make it more difficult to get to the good ending. Basically give them another choice to make on those paths, one that seems costly, so it is unlikely they will take it, but IF THEY DO then they are back on their way to the good path (whether they know that or not).

Then you have an "easy" way to win, and a few "difficult" ways to win.

  • 1
    If only they could make games like that. But sadly the software industry hasn't solved the n-body problem either. I remember when Fable first came out. Everything was supposed to effect everything.... Yeah, it's not so much like that. Every contingency would require you to rewrite the story from then on. And for every contingency therein, repeat.
    – Mazura
    Commented Feb 9, 2018 at 20:13
  • @Mazura I don't think that's true (I have a PhD in CS and have been a professional programmer over 40 years). You choose to fight the monster with poison instead of a spell; the poison doesn't work, the monster chases you, you cross the river and it won't. If you go right, you will lose. If you go left, which seems to be away from your goal, you will find a bridge back across the river, and an amulet that protects you from the monster. Walk back. The monster runs away from you, and you end up on the same path as if you had killed it. It is just a side-loop in the path diagram.
    – Amadeus
    Commented Feb 9, 2018 at 20:37
  • Agreed. @Mazura would be correct if every choice the player could possibly make could effect the outcome, but you're only talking specific choices. You're not opening this to "I pick up a pine cone and eat it and then..." I believe that you are only suggesting a limited number of change points.
    – NomadMaker
    Commented Feb 10, 2018 at 9:43
  • @NomadMaker Exactly, add loops to a third of the choices. More like life, not every bad choice is fatal. Some stupid decisions just ruin the car and cost us a month's pay, and a week taking the bus to work. Time and fortune. e.g. Break the game into 'Stages'. Wander around Stage 1 all you want, the only way forward is through the door guarded by the despondent troll, there are three ways: bring him flowers for his favor (easy), bring the amulet to scare him (easy to find but expensive), kill him in battle: A lethal choice without the rusty iron dagger, which is in a remote corner of Stage 1.
    – Amadeus
    Commented Feb 10, 2018 at 11:03

The first thing that came to my mind when I read your question was the Black Mirror Netflix series. It has very high ratings despite the dark tone and the fact that few episodes actually have a happy ending. Even the happy endings are seldom purely happy, often they just mean that it wasn't the worst possible outcome for everyone. It's really hard to describe the series without spoiling it, so I recommend you just watch it, reflect it, and learn from it. The point is, you can absolutely tell a compelling story without a happy ending, but the reader should probably somehow be made aware in advance that that's the name of the game.

Next I thought of some games that would match the situation. The first one I thought of was 80 Days, a work of interactive fiction where you are supposed to circumnavigate the world in 80 days or less. If you don't make it you've lost the game, but you don't feel too bad about it because the characters just go like "Oh well, it was an interesting adventure anyway, and we've got a lot more to come!" or something like that.

But more importantly, when you play the game, you learn about the world, and that knowledge carries with you to the next playthrough. So if during your first game you learn that there's a way to travel from, say, Budapest to Kiev, then in the next game you know there will be such a connection right when you start the game, and this time you don't need to buy a train map or listen to some gossip or anything like that to find that out. And even if you know about a connection, it might not be as useful a piece of information as it was in the previous game, due to random events or other kind of change of plans.

Finally, I thought about Nethack. It's a roguelike where the default expectation is that you will die, and there are dozens of ways to die. People are known to have played it for decades already without ever completing it. But even when your character dies in the dungeon, you will have learned something that you can use in your next game. What's more, dead characters may actually appear in the future games as ghosts of themselves! Then you might be forced to fight a previous version of yourself, or, in case of online servers, the ghost of some other poor player. And should you defeat your ghost, you'll get (at least some of) the items you were carrying when your game ended. So this way your previous losses may again help you in your current game.

Anyway, how long are you planning for one playthrough to last? Because I can tell you I feel a lot worse if I spend 30 hours on a game just to be told "Everybody died, game over, you lost", versus spending, say, just three hours and encountering this ending. I love the Civilization series, but when I lose after a loooong playthrough, I don't feel like starting a new game any time soon. So definitely plan and balance the play time carefully. Don't go for anything too long with the format that you've chosen.

And if you can add any upside to the bad endings at all, I think that would be good. (The Old One didn't enslave the entire world, only the Americas! The whole family didn't die, one of the children only lost their legs! Etc...)


It's a bad design choice that is unfortunately prevalent in video games. Personally, I think that even the idea of a "canon" ending defeats the purpose of having a branching narrative at all.

But to be clear: The problem is not that the endings are bad, the problem is that they are not satisfying. In a story where you can affect the outcome, the only acceptable ending for a player is the one that is satisfying on some level. Ideally, your endings are satisfying in different ways. And ideally, you will have a specific unsatisfying ending for each satisfying ending, as a counterpoint, one that is actually supposed to be unsatisfying (similar to the Game Over screen - it's something that the player is NOT supposed to accept, but urges them to try again).

I'll give a simple example: a dating sim. There are five different love interests. Only one is considered canon. If you end up with him, you get the good ending where you are happy ever after. If you choose one of the others (or rather, your choices within the game lead you to one of the others), then you get a bad ending where he turns out to be the wrong choice.

That sucks. It's bad design. All five love interests should be a valid choice, because it's a choice the player makes on personal preference. You cannot dictate whom they find hot and likable. What you should do instead is make good AND bad endings for each of the love interests. So you can screw it up with each of them, but you can also have a satisfying ending with each of them.

It's also a bad idea to have levels of success, e.g. if you have five characters that need to be saved at some point, and the ending changes in dependence of how many you saved. No matter the outcome, the only satisfying ending will be the one where you save all of them. However, if exactly two of them die no matter what the player does, but they can affect which ones die, then they accept the outcome and the narrative they chose is satisfying to them.

A good example is Dishonored, in my opinion. They have two different endgames, one that is very bright and positive and one that is dark and depressing. Both are awesome. It depends on what you are going for personally. If you decide to be the vengeful assassin, you get an endgame befitting the dark tale of a vengeful assassin. The showdown is on a dark rainy day. If you decide to be the honorable agent of justice, you get an appropriate endgame, where the showdown is on a sunny day. You can still screw it up in both cases, but if you do it right, the dark ending is still satisfying because it offers what the player was going for.


The game "Marvel: Ultimate Alliance" had a system where each side quest's outcome (and some choices that were tied to the main plot and couldn't be avoided) were recorded. At the end of a full play through, the marvel character Uatu tells you how each of these choices (including never acting on the choice) will alter the course of history. Most of these choices have a good and bad outcome that ties the good outcome to completion of the side quest and the bad outcome to non-completion. In the case of the unavoidable mission, there is no correct answer and both choices receive a negative outcome where the X-Men are disbanded... though one is with a far lower body count.


I know this is old but I feel compelled to add an answer that I don't see.

Many stories have an external plot and an internal plot that's going on inside of the character.

  • The external plot is obvious: slay the dragon, or escape from the murderer, or win the championship.
  • The internal plot is related to the protagonist's flaw. Maybe they're selfish and they need to learn to be selfless, or they're cynical and they need to learn to enjoy life, etc.

Oftentimes the best stories have the external plot pitted against the internal plot.

It can be of great literary value to have one ending where the character gets the (external) thing they want, but fail to make the internal change they need to, and another ending where they give up the external thing they want, but make the internal change they need to.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.